Tell the agent what to change. Review the diff.

What it is

The panel sits on the right of the editor and is open by default. Its header reads “AI chat”, with “working on ‘<document>’” beneath it, so you always know which document a request lands on. Ask a question and you get prose, with step references by line number. Ask for a change and you get a proposal headed “Here’s the change — nothing is applied until you say so:” — a compact diff with +added and −removed counts, the changed lines with one line of context, and Apply and Discard buttons underneath. Long diffs fold after 14 rows and say how many more rows the document has.

The agent never writes on its own. A proposal waits in the conversation until you decide, and the document you are editing stays exactly as it was while you read it.

How a proposal is built

Every request carries the current document (the first 16,000 characters of it), the recent conversation, and summaries of older stretches. The model replies in one of three modes: EDIT returns the complete revised document, NEWDOC returns a separate new document, ANSWER returns prose. Runic diffs an EDIT against your text on the client, so what you review is the real change, not a description of it.

For process documents the reply goes through the same parser and structural audit the editor uses before you see it. If the structure is broken — a syntax error, a one-branch decision, an unreachable step, an empty branch — the model gets one correction pass, and the fix is kept only when it scores measurably better. When the correction was kept, the reply carries a small “self-checked” tag under the message, next to the model label.

One model per job

Ordinary turns run on Claude Sonnet 5. A request is routed to Claude Opus 5 when it is whole-document work: it carries an attachment, it contains a whole-flow word such as redesign, rewrite, restructure, from scratch, overhaul, audit, architecture, end-to-end or “design a” (English or Turkish), it is longer than 600 characters, or it asks to fix, improve or clean a document over 80 lines. The label under each reply tells you which model answered.

Questions that need outside knowledge — regulations, GDPR or KVKK, ISO standards, industry best practice, benchmarks — switch on web search for that call, capped at three searches. An answer that used it is tagged “web research” and is asked to end with a “Sources:” line. Document bodies never contain links or citations.

Apply is guarded

Apply only works while the proposal’s document is open and still equals the text the proposal was computed against. Edit anything in between, or switch documents, and the button greys out with the note “document changed”; undo the edit and it lights up again. A proposal that is still pending when its messages are archived, or whose document has been deleted, is marked “Superseded — the document moved on” instead of being forced in. On a team document where you are a viewer, the agent tells you to ask an editor.

When you do apply, the change types itself in from the first line that differs, so the diagram redraws step by step rather than snapping; a proposal applied to a document that is not open is written in one go. The entry flips to “Applied ✓” and gains a ↩ Restore button — the checkpoint.

Limits

The chat needs a signed-in account with a verified email. Each account gets 60 chat messages an hour. The server reads the first 2,000 characters of a message and the first 16,000 of the document; on a longer document a whole-document edit will show the tail as removed in the diff, so check before applying. Chat works on sequence diagrams too, but the audit button, the diagram preview, the selection chip and the self-check pass are process-only. Fullscreen on the diagram hides the panel along with the editor; Esc brings them back.

How to use it

1. 1
   Open the panel

   It is open by default on the right. If it is collapsed, click the ✨ rail on the right edge (tooltip “Open AI chat”) or “AI chat” in the top bar.
2. 2
   Say what you want

   Type in the box labelled “Ask, or describe a change…” and press Enter. Shift+Enter adds a line; the Send button does the same as Enter.
3. 3
   Read the reply

   A question gets prose. A change gets a diff with +/− counts. The model label (SONNET or OPUS) and any “web research” or “self-checked” tags sit under the message.
4. 4
   Apply or Discard

   Click Apply to write the change in, or Discard to drop it. While a proposal is pending the entry says “previewing in the diagram →” and the flow pane shows it.
5. 5
   Come back later

   Every applied entry keeps a ↩ Restore button that returns the document to how it was before that change.
6. 6
   Start fresh when you want

   “Clear” in the panel header wipes the conversation, including archived stretches and their checkpoints.

Commands & shortcuts

Tips

• Say “do what you suggested” — the agent builds the edit from the conversation instead of asking you to repeat it.
• Use the words “redesign” or “from scratch” when you want a whole-document rewrite; that routes the turn to Opus.
• Ask “what does GDPR require here?” style questions to get a researched answer with a Sources line.
• For a change to a few specific lines, select them and use Copy to chat — the edit is then confined to that region.
• If Apply is greyed out with “document changed”, ask again; the new proposal is computed against the current text.

Questions